AI Image Edit: open-source mobile-friendly UI for image editing (ComfyUI backend, RunPod template included) Developer pulb released AI Image Edit, an open-source, web-based image editing and generation app that runs on a ComfyUI backend and is designed for mobile use. The app supports inpainting with a mask editor and soft blending, colored annotations to direct model edits, chained step-by-step edits, a before/after slider, aspect ratio presets and optional LoRAs, and ships with Qwen-Image 2.1 (full or quantized GGUF), an official 8-step Turbo version and Qwen-Image-Edit 2511 with LoRAs, plus support for custom workflows. It deploys via a RunPod template, Docker or a Hugging Face Docker Space, generating roughly 10 s per 1K image on an H200 and about 31 s on a 24 GB RTX 4090 with the GGUF model, keeping images and prompts on the user's own hardware or a disposable container. Hi everyone, I built AI Image Edit, an open-source app for AI-assisted image editing and generation. The UI is web-based and designed to work well on phones and sits in front of interchangeable model workflows running on ComfyUI. What it does - Inpainting with a mask editor and soft blending - Colored annotations drawn on the image to point the model at what to change - Feeding any result back in as the next input for step-by-step edits - Before/after comparison slider - Aspect ratio presets and optional LoRAs Models - Qwen-Image 2.1, in full quality or as a quantized GGUF variant for GPUs with less memory - An official 8-step Turbo version for much faster generation - Qwen-Image-Edit 2511 with LoRAs for quick edits - You can also add your own workflows Try it - RunPod template recommended , Docker, or a Hugging Face Docker Space - Roughly 10 s per 1K image on an H200, and about 31 s on a 24 GB RTX 4090 with the GGUF model It runs on your own hardware or in a disposable container, so your images and prompts stay with you. GitHub: GitHub - pulb/ai image edit: Deployable AI Image Editor · GitHub https://github.com/pulb/ai image edit Feedback, bug reports and ideas are very welcome. It’s a hobby project, so I’m happy to hear what’s missing.